Deep Brain Stimulation for Parkinson's Disease PDF

111 111 Deep Brain Stimulation for Parkinson Disease Fedor E. Panov, Andrew K.P. Conner, Paul Larson, Alastair Martin, and Ph...

111 111 Deep Brain Stimulation for Parkinson Disease Fedor E. Panov, Andrew K.P. Conner, Paul Larson, Alastair Martin, and Philip Starr Anticholinergic treatment was used as early as the mid-19th This chapter includes an accompanying lecture presentation that century. Arvid Carlsson and George Cotzias introduced oral has been prepared by the authors: Video 111.1. levodopa/carbidopa as the “gold standard” of medical therapy in 1968.8 Surgical treatment of PD began in the 1940s with resection of premotor and motor cortices in hopes of alleviating parkinsonian KEY CONCEPTS tremor.9 The resultant improvement had to be weighed against significant iatrogenic motor deficits, while no effect was seen on Deep brain stimulation (DBS) is the dominant surgical either rigidity or bradykinesia.9 The move toward basal ganglia intervention for Parkinson disease (PD) because of its and thalamic targets addressed these shortcomings. Interruption reversibility, adjustability, and applicability for bilateral of the pallidofugal fibers exiting the GPi by Spiegel and intervention. Wycis10,11 and by Meyers12 improved both tremor and rigidity. In 1952, Irving Cooper’s fortuitous sacrifice of the anterior DBS is thought to work by reducing abnormally choroidal artery in a 39-year-old man incapacitated by tremor elevated neuronal synchronization in the basal ganglia- and bradykinesia alleviated the patient’s symptoms without thalamocortical motor loop. motor or sensory deficits.13 DBS can improve bradykinesia, rigidity, and tremor, Lesioning work on the basal ganglia and thalamus largely and can do so with less fluctuation in motor signs than ceased in the 1970s because of the dramatic immediate effects antiparkinsonian medications. of oral levodopa/carbidopa and the significant risks of surgery. The subthalamic nucleus (STN) and the pallidum DBS By the 1990s, however, recognition of the long-term side effects targets have equivalent efficacy; however, targeting of the medication—dyskinesias and motor fluctuations—brought the STN allows reduction of levodopa more than does about a renewed interest in surgical solutions.14,15 targeting its counterpart. On the other hand, placement At the same time, the integration of modern imaging in the STN is contraindicated in patients with significant techniques such as CT in the late 1970s and MRI with frame- underlying psychiatric illness. based stereotaxy in the 1980s improved the safety and accuracy of surgery at deep brain targets. Attempts to use DBS in thalamic, DBS achieved with implantation by means of the basal ganglia, and cerebellar regions for movement disorders microelectrode recording guided or “awake” paradigm were made in the early 1980s.16,17 has shown equivalency in patient outcomes when In 1987, Benabid and colleagues showed that high-frequency compared with the interventional MRI or “asleep” stimulation could mimic a lesion in a controllable, reversible paradigm. manner.18 Yet stimulation of the ventralis intermedius thalamic nucleus target again aided only the tremor symptomatology, leaving rigidity and bradykinesia untreated. Series of safe and Parkinson disease (PD), a progressive synucleinopathy of effective pallidotomies19 from the early 1990s resurrected the unknown origin, is the second most common neurodegenerative concept of GPi lesioning for those symptoms, yet were technically condition behind only Alzheimer disease, affecting 1% of the difficult procedures limited to one hemisphere. GPi-DBS was population older than 65 years.1,2 With the prevalence doubling introduced in 199420 as a safer reversible alternative with an per decade of life, the number of patients is predicted to increase ability to implant and modulate both hemispheres for bilateral as the world population ages.2 and axial symptomatology. The goal of this chapter is to provide an overview of the In 1990, Bergman and associates21 showed in a nonhuman treatment of the motor abnormalities of PD using deep brain primate model of PD that the induction of parkinsonism is stimulation (DBS) with emphasis on two targets: the subthalamic associated with excessive and abnormally patterned discharge nucleus (STN) and the globus pallidus interna (GPi). Two surgical in the STN, and that ablation of the nucleus alleviated all methods preferred at our institution are highlighted: surgery parkinsonian motor signs. Based on this work, Limousin and guided by microelectrode recordings (MERs) and test stimulation coworkers implanted the first chronic subthalamic stimulator for in the awake state, and surgery guided by interventional MRI PD in the early 1990s22 and subsequently documented alleviation (iMRI) under general anesthesia, without physiologic testing. of all cardinal motor signs of PD in a case series in 1998.23 BRIEF HISTORICAL OVERVIEW ANATOMY AND PHYSIOLOGY OF TARGETS Descriptions of PD symptoms and treatment with the dopami- Both the STN and GPi are components of the basal ganglia, a nergic extract “cowage,” from the Mucuna pruriens seed, appear collection of subcortical nuclei involved in scaling and focus- in the Ayurveda, an ancient medical text of the Indian subcon- ing of movement, as well as motor learning. The basal ganglia tinent from 1000 bce.3,4 The Nei Jing, a 2500-year-old medi- also include the striatum (caudate and putamen), globus pallidus cal text from China, echoes similar ideas,5 yet the first western externa (GPe), and substantia nigra, subdivided into the sub- description of the disorder emerged only two centuries ago when stantia nigra pars reticularis and substantia nigra pars compacta. a Londoner, James Parkinson, described six patients with “paraly- DBS for treatment of PD motor symptomatology is based on the sis agitans.”6 Sixty years later, Jean-Martin Charcot refined the “segregated circuit hypothesis.”24 The numerous functions of earlier description and coined the term “maladie de Parkinson.”7 the basal ganglia within the cortex–basal ganglia–thalamus loop 731 732 SECTION 5 Functional Neurosurgery (motor, oculomotor, associative, and limbic) run in parallel and motor (M1) and sensory cortices during postural maintenance occupy anatomically distinct areas of the nuclei. It is therefore in but is superseded by the high-frequency band (76–100 Hz) with principle possible to target the motor areas without compromis- planning and execution of movement.32 The basal ganglia show ing nonmotor functions. an increase in neuronal synchronization in the beta activity in The STN, shaped like a small, thick biconvex lens, is located PD, in the form of abnormally long “bursts” of beta activity.33 medial to the internal capsule, lateral to the red nucleus, superior DBS at both the STN and GPi may exert its therapeutic effect by to the substantia nigra, and inferior to the thalamus. The GPi is desynchronization of neuronal activity at specific frequencies.34 located medial and inferior to the medial medullary lamina and GPe. The GPi overlies the choroidal fissure and the optic tract. Medially, it is limited by the genu and the lateral aspect of the PATIENT SELECTION posterior limb of the internal capsule. The most important part of DBS for PD is patient selection In the basal ganglia–thalamocortical circuit (Fig. 111.1), the (Table 111.1). The key to proper patient selection is a combined major input structures to the basal ganglia are the striatum and approach that includes evaluation by specialists in movement dis- STN, and the GPi serves as the main output. The input and orders, neurology, neurosurgery, and neuropsychology. An inter- the output nuclei are connected by a direct massive GABAergic disciplinary clinic in which these specialties provide integrated striatopallidal pathway, and by an indirect route via the GPe and care streamlines the process of patient selection and work-up. STN before arriving at the output nucleus, the GPi. The balance between the activating direct and the inhibitory indirect pathways controls movement. In the “rate model” developed originally by Indications Albin, Young, and DeLong, the parkinsonian state is modeled as A clear diagnosis of idiopathic PD by a movement disorder a hyperexcitation of the STN causing an imbalance in favor of the neurologist is paramount because many forms of atypical par- indirect pathway and excessive GPi excitation.21,25 Targeting the kinsonism, such as progressive supranuclear palsy and multiple STN or GPi with lesioning or stimulation was thought to correct system atrophy, may resemble PD but do not respond to sur- excessive and abnormally patterned basal ganglia output. Some gical treatment. The patient’s examination off medication and elements of the rate model have been confirmed experimentally. subsequent improvement with a supratherapeutic oral levodopa Elevated GPi and STN discharge rates in PD, compared with dose, called “on-off” testing, is a strong predictor of the quantita- nonparkinsonian conditions, were seen by Starr and colleagues,26 tive improvement in scores on Part III (motor subscale) of the Schrock and associates,27 and Steigerwald and colleagues.28 Unified Parkinson’s Disease Rating Scale (UPDRS III) that can Optogenetic confirmation in rodents of the prokinetic and be expected from bilateral DBS. The two indications for surgery antikinetic functions of the direct and indirect pathways, are motor complications from long-term medical therapy and/ respectively, was shown by Kravitz and coworkers in 2010.29 The or medically intractable tremor. The most common motor com- model does not, however, explain the benefit of pallidal stimulation plications are dyskinesias and motor fluctuations (rapid, unpre- in hyperkinetic disorders. Its initial assumption that STN-DBS dictable cycling between effectively medicated and inadequately suppressed downstream basal ganglia activity has also been called medicated states).35 Postoperative reduction in the severity, dura- into question.30,31 The rate model remains important because of its tion, and frequency of “off” periods, as well as reduction of med- heuristic value, as newer theories aim to rectify its shortcomings. ication-induced complications, improves quality of life. DBS for The most influential contemporary model of brain network early motor complications of PD is now being investigated.36,37 dysfunction in PD is the “oscillatory synchrony model,” which posits that akinesia and bradykinesia in PD are due to excessive neuronal synchronization in specific frequency bands, especially Contraindications the beta band.1 The beta band (13–35 Hz) is present in primary Although most patients with PD have at least some cognitive impairment that can be detected with neuropsychological test- ing, severe cognitive impairment is a contraindication for DBS because dementia may worsen as a result of surgical intervention, Cortex Cortex Cortex and severe cognitive impairment can supersede motor impair- ment as the primary driver of disability. A marked cognitive dys- Putamen Putamen Putamen function documented on the Mini Mental State Examination38 or a low score on the Mattis Dementia Rating Scale diminishes the DBS quality-of-life benefit and serves as a relative contraindica- SNc SNc SNc tion to DBS in PD.39,40 Improved motor function in a severely demented patient can be dangerous because increased mobility GPe GPe GPe VL VL VL TABLE 111.1 Response of Parkinson Disease (PD) Symptoms to Deep STN STN STN Brain Stimulation (DBS) DBS Improves DBS Unlikely to Benefit GPi GPi GPi A B C Tremor Autonomic function (constipation, Dyskinesias poor temperature regulation, Figure 111.1. The classic rate model of the basal ganglia– Rigidity orthostatic hypotension) thalamocortical circuit. (A) Normal state showing excitatory Motor fluctuations Cognition projections (open arrows), inhibition (filled arrows), and relative activity Bradykinesia and “off”-period gait Hypophonia (arrow width). (B) Parkinsonian state. (C) Effects of subthalamic nucleus freezing Postural instability and “on”-period Classic idiopathic PD gait freezing (STN) lesioning on the circuit. GPe, Globus pallidus externa; GPi, Mood depression or anxiety globus pallidus interna; SNc, substantia nigra pars compacta; VL, Atypical parkinsonisms (PSP, MSA) ventrolateral nucleus of the thalamus. (From Bergman H, Wichmann T, DeLong MR. Reversal of experimental parkinsonism by lesions of the MSA, Multiple system atrophy; PSP, progressive supranuclear palsy. subthalamic nucleus. Science. 1990;249:1436–1438.)  CHAPTER 111 Deep Brain Stimulation for Parkinson Disease 733 may lead to falls. The danger may be exacerbated by a potentially its greater anatomic variability and a relative lack of consensus on increased impulsivity caused by STN-DBS.41,42 the subregion of this relatively large structure that is optimal for 111 Age remains an independent prognostic factor in a majority DBS therapy. of surgical fields. As a disease affecting 1% of patients older than 65 years, PD carries with it the comorbidities of advancing age. Hypertension, diabetes, and need for anticoagulation all increase Alternative Target: Ventralis Intermedius Nucleus surgical risk. Age and hypertension increase intraoperative The first target used historically for PD, the ventralis intermedius hemorrhage rates in DBS.43,44 Although each candidate must be nucleus of the thalamus, results in only reduction of tremor, but reviewed on an individual basis, the quality-of-life improvement remains of some value in tremor-predominant PD.56 Because the seen in younger patients undergoing DBS for PD diminishes in STN and GPi alleviate tremor together with other motor symp- the group older than 65 years.45 toms, most practitioners favor those targets over the ventralis Patients with poor axial scores on the UPDRS III combined intermedius nucleus. with severe postural instability tend to not benefit from DBS. Hypophonia, another levodopa-nonresponsive symptom, is Unilateral Versus Bilateral Versus Staged unlikely to improve with stimulation. Patients and families should be warned of potential worsening in voice strength Stimulator Implantation postoperatively. Unilateral implantation should be considered if there is gross asymmetry in the symptoms. Advanced age or presence of preop- erative cognitive deficits may make unilateral surgery more pru- TARGET SELECTION FACTORS dent, with a faster recovery time because of decreased operating time and limitation of postoperative edema to one hemisphere. It Patient Symptomatology is sensible to alleviate the symptoms on the severely affected side Several large randomized studies that have compared STN and with a unilateral implant first and, if required, perform implanta- GPi as targets showed no significant difference in improvement of tion on the other side at a later date. Even unilateral STN-DBS motor symptoms after DBS.46-48 Mood and cognition seem to be to treat the more severely affected side will significantly improve at slightly higher risk for decline following STN-DBS46,48 (Table motor function. Unilateral DBS may decrease some aspects of cog- 111.2). Patients with borderline cognitive function seem to main- nitive performance,57 but less so than simultaneous bilateral DBS. tain their function better, and their quality of life improves more with GPi-DBS.47 Decision making under stress can be altered by STN-DBS, increasing impulsive behavior and errors in judg- SELECTION OF SURGICAL TECHNIQUE ment.41 This argues in favor of using GPi-DBS for cognitively At the time of its introduction, DBS was performed using a com- compromised patients whose cognitive deficit is not so severe as bination of anatomic and physiologic targeting. In the traditional to preclude surgical intervention. technique, frame-based stereotaxy uses preoperatively acquired Swallowing function deteriorates with PD and can lead to magnetic resonance images to define a “starting point” for tar- aspiration pneumonia and death. A recent retrospective review get localization, and this initial anatomic target is refined or con- suggests that, unlike GPi-DBS, STN stimulation can worsen firmed with microelectrode recordings (MERs) performed with swallowing.49 Weight gain is seen with STN-DBS, whereas the patient awake. The use of MER increases intraoperative time, results are inconclusive with GPi stimulation.50-52 Long-term and its application in certain populations (those with significant benefits in favor of STN-DBS include possibly increased battery anxiety, advanced age, or severe “off”-period pain) may be prob- life (because of lower voltage settings) and greater reduction of lematic. It has been suggested that MER increases the risk of dopaminergic medication. Only one of four randomized studies of DBS,43,58 but this additional risk, if it exists, is small. GPi-DBS versus STN-DBS has shown better motor function in Recent advances in the technical approach to DBS include the “off” state following STN-DBS compared with GPi-DBS.40 novel skull-mounted aiming systems, the use of “frameless stereotaxy,” and “asleep” DBS with iMRI or CT used to place leads using anatomic guidance only. The microTargeting Surgeon’s Experience platform stereotaxy system (FHC) was introduced over a decade The experience of the surgeon remains of vital importance.53,54 ago as an alternative to frame-based systems for awake, MER- As with any procedure, a learning curve exists for both targets. guided surgery and was reported on by Konrad and colleagues.59 Benabid and associates55 reported an improvement in complica- Neuronavigation-guided awake DBS using the Nexframe tion-free surgical rate from 37.3% in their first 150 bilateral pro- (Medtronic) skull-mounted aiming device was established by cedures to 72.7% in the next 150. Seijo and coworkers reported Henderson and colleagues.60 Our team showed accurate STN that the rate of significant adverse events in their series dropped and GPi lead placement using a frameless, asleep iMRI technique from 14.6% in the first 7 years to 8.8% in the last 7 years.44 If true with outcomes comparable to those of traditional frame-based, equipoise as to target selection exists after an exhaustive work-up, MER-guided placements.61,62 Burchiel and associates used the surgeon will likely choose the procedure with which he or she intraoperative CT-guided navigation without MER for accurate is most experienced. With deference to personal experience, GPi lead placement.63 Numerous additional innovations in targeting is viewed by many to be the more challenging target because of will likely occur in the next decade. The dichotomy of awake versus asleep surgery remains a major choice in procedure selection. Recent studies have suggested that TABLE 111.2 Target Selection Based on Treatment Goals accuracy of placement and complication profiles are similar,60,61,63 yet the patient experience differs. In our practice, younger Subthalamic Nucleus Globus Pallidus Interna patients or patients without severe tremor or painful “off”-period Greater decrease in levodopa Lower chance of cognitive decline dystonia tolerate awake procedures well. Older patients with requirements Easier postoperative programming significant tremor, anxiety, and major off-medication disability Easier surgical targeting Maintenance of letter verbal are better suited for the asleep iMRI approach. Although some Weight gain fluency physiologic recording can be done under general anesthesia with Possibly longer battery life Mood stability frame-based surgery in a traditional operating room, the quality of information obtained is much less than in awake patients. 734 SECTION 5 Functional Neurosurgery and 12 mm lateral to the midcommissural point. The T2-FSE AWAKE SURGICAL TECHNIQUE image set is then used to adjust the target with respect to the unique There is no consensus on specific methods for performing DBS anatomy of each patient. The anterior border of the red nucleus is in the awake state. Our methods, as well as those of other centers, visualized on an axial slice 4 mm below the midcommissural plane. continue to evolve with technologic improvement and experi- A line drawn tangential to the anterior apex of the red nucleus is ence. We present an approach that has resulted in good outcomes seen cutting through the STN (Fig. 111.2). We pick the target and low complication rates at our institution. point 2 mm lateral to the medial border of the STN along this line. Alternatively, if the lateral border of the STN is also well visualized, we target the middle of the nucleus along the line. Preoperative Imaging Three MRI sets performed within a week of surgery are used for Globus Pallidus Interna. The approximate coordinates used anatomic targeting. A gadolinium-enhanced volumetric three- for initial GPi targeting are as follows: 2 mm anterior, 5 mm dimensional T1 data set covering the whole brain in 1-mm axial inferior, and 21 mm lateral to the midcommissural point. The cuts is used together with two two-dimensional image sets opti- T2-FSE or FSE-IR image set is then used to adjust the target, mized for the best delineation of the targets: an axial fast spin accounting for individual patient variability, which is high for this echo inversion recovery (FSE-IR) sequence optimized for pallidal target. We measure the pallidocapsular border on the axial slice targeting,64 and an axial T2-weighted fast spin echo (T2-FSE) at the level of the anterior commissure–posterior commissure sequence for STN targeting, acquired as interleaved sequences to (AC-PC) plane and divide it into thirds. The target is chosen by provide contiguous slices (zero interspace). Images are imported drawing a 3- to 4-mm line perpendicular to the pallidocapsular into a stereotactic surgical planning software package (e.g., border at the junction of its posterior one-third and anterior two- STEALTH Cranial or BRAINLab) for planning. thirds65 (Fig. 111.3). Trajectory. The approximate initial trajectory for both STN Operation and GPi stimulation is 60 degrees from the AC-PC line in the We place a standard stereotactic Leksell head frame (Elekta) in sagittal plane and 0 to 15 degrees from the vertical in the coronal the preoperative holding area with conscious sedation. Mouth plane. Patient-specific adjustments include avoiding cortical and eyes remain free for airway access and continuing neurologi- sulci and vascular structures superficially and deep. If the lateral cal examination. A CT scan performed after the frame placement ventricle is crossed along the trajectory, we adjust the entry point with the fiducial box is computationally fused with the preopera- because ventricular violation is shown to increase morbidity66 and tive magnetic resonance images to serve as image guidance for may reduce the accuracy of placement. Navigation views are very the procedure. After the fiducial marker registration, the image helpful in entry point adjustment. sets are reformatted to produce images orthogonal to the mid- commissural plane. Positioning and Exposure The patient is placed in a semisitting position, and the Mayfield Targeting headrest is used to fix the head frame in neutral position to the Subthalamic Nucleus. The approximate coordinates for locali- operating table. To avoid undue stress on the neck, the bed zation of the STN are as follows: 3 mm posterior, 4 mm inferior, position should be finalized prior to fixating the head. A two- dimensional or three-dimensional fluoroscopy unit is set up with the head correctly centered. The head is prepped and draped to ensure accessibility to the face from the unsterile side. After a Ventral skin incision just posterior to the planned skull entry location, STN Putamen RN GPe GPi SC Dorsal Figure 111.2. Midbrain at 4 mm below the midcommissural Figure 111.3. The axial cut at the midcommissural plane. The plane. The line parallel to the ventral border of the red nucleus (RN; medial border of the globus pallidus interna (GPi) is split into thirds dashed line) is extended to the midpoint of the subthalamic nucleus (dashed line,) and a point 3 to 4 mm lateral and perpendicular to the (STN; black dot) or, if the lateral border is difficult to interpret, at least 2 line at the junction of the posterior one-third and anterior two-thirds mm away from the STN medial border (short vertical bar) for targeting. (short vertical bar with black dot) is used for targeting. GPe, Globus SC, Superior colliculi. pallidus externa.  CHAPTER 111 Deep Brain Stimulation for Parkinson Disease 735 the bur hole is made with a 6-mm cutting drill. A recess is drilled recorded simultaneously. The inferior border of the STN is around the bur hole to lower the profile of the lead anchoring signaled by a sudden decrease in the background noise, and the 111 device, which aids in cosmesis and wound healing (Fig. 111.4). entrance into the substantia nigra pars reticularis is signaled by an In rapid sequence the dura is opened, hemostasis is achieved, and increase in the discharge rate to 50 to 70 Hz. While in the STN, the microelectrode guide tube is advanced past the pial surface. audible modulation of neuronal discharge by passive contralateral This allows the bur hole to be quickly filled with fibrin glue to joint movement confirms localization of the motor area of the decrease pneumocephalus and chances of air emboli. nucleus, which is dorsal, posterior, and lateral to nonmotor areas. Microelectrode passes are always sequential, with the location of each pass informed by the recordings gathered during Microelectrode Recording previous passes. Good initial recording, presence of passive joint MER is a technique with very high spatial resolution, thus aug- movement responses, and a span of STN cells over a distance of menting preoperative imaging, which may lose accuracy because 5 mm obviate the need for further passes. of nonlinear image distortions, mechanical inaccuracies of the In the GPi (Fig. 111.5B), recordings begin 15 mm dorsal to frame, and unpredictable brain shift.67 We use high-impedance the anatomic target. We usually record the striatum followed by (0.3–0.8 MΩ at 1000 Hz) tungsten or platinum-iridium micro- the GPe before reaching our target. The initial low-frequency electrodes for recording of single-unit extracellular action poten- striatal discharges (0–10 Hz) increase to 30 to 60 Hz on GPe tials.68 Controlled advance of the microelectrodes and DBS entrance. Frequent “burster” and “pauser” (Fig. 111.5B) electrodes is performed with a micropositioner (NeuroNav discharge patterns are seen. When entering the GPi, we see a Drive, Alpha Omega; Elekta micropositioner). The low deform- further rate increase to 60 to 100 Hz. The lamina separating the ability of the microelectrode in brain tissue allows for an accurate GPi and GPe contains “border” cells with a fairly regular 20- to unsupported trajectory of up to 30 mm. The relatively higher 40-Hz discharge pattern. The somatotopic organization of the deformability of the DBS electrode mandates more guidance GPi (lower extremity motor-responsive cells medial and dorsal from a rigid guide tube, no farther than 15 mm from the target. to upper extremity cells) is useful in checking the laterality of the For the STN (Fig. 111.5A), the recordings typically begin 10 recording, based on movement-related modulation of neuronal mm dorsal to the anatomic target. On entrance into the STN, discharge. The recordings are stopped 1 to 2 mm after the last neuronal action potentials are detected at firing rates of 20 to GPi unit to minimize chances of vascular injury in the choroid 50 Hz. Because of high cell density, multiple units are often fissure. Microstimulation or light-evoked fiber activity can be used to identify the optic tract below the base of the pallidum. Deep Brain Stimulation Lead Implantation and Macrostimulation Several new DBS hardware platforms have become available in the last 3 years. Our current technique is based on DBS hardware from Medtronic or Boston Scientific. With use of a rigid guide tube no farther than 15 mm above target, the lead is placed with the bottom of its distal contact at the microelectrode-recorded ventral limit of the target for the GPi. For the STN, the lead is positioned so that the middle of the contact area is within the dorsal STN. The final position check includes macrostimulation between the dorsal and ventral contacts of a 7.5-mm array, at a rate of 180 Hz, with a 60-microsecond pulse width. The voltage or current is increased to assess the threshold for stimulation- Figure 111.4. Countersinking the bur hole. induced adverse effects and that for stimulation-induced changes, including tremor, rigidity, and bradykinesia reduction. With STN implantation, the voltage threshold required to STN region GPi region activate the medial lemniscus, posteromedial to the target, and the corticobulbar tract, lateral to the target, indicates the lead’s Striatum GPe bursting cell proximity to these structures. Stimulation-induced adverse effects should occur in the 6- to 10-V or 4- to 6-mA range. If strong persistent paresthesias from medial lemniscus stimulation, Dorsal GPe pausing cell or dysarthria and facial contractions from corticobulbar tract thalamus stimulation, occur below 2 V or 1.5 mA, the lead location may be suboptimal and may need to be revised. Contralateral gaze STN “Border” cell deviation, from activation of the frontopontine bundle anterior to the STN, may be observed above 8 V or 6 mA. A lead placement error in the anterior, medial, and ventral direction may cause SNr GPi ipsilateral monocular adduction with current spread to the nucleus or tract of the third cranial nerve. Acute stimulation- A B induced mood changes may be related to stimulation of the limbic Figure 111.5. Microelectrode recordings of spontaneous neuronal (anteromedial) portion of the STN. If no effects are seen at 10 V activity from the thalamus and basal ganglia in the (A) subthalamic or 6 mA, the pulse width can be increased to 200 microseconds. nucleus (STN) and (B) globus pallidus interna (GPi) regions. Each The likely reason for lack of effects after such an adjustment is an trace is a 1-second recording from a Parkinson disease patient. GPe, electrical malfunction or misplacement of the lead dorsally. Globus pallidus externa; SNr, substantia nigra pars reticularis. (Modified With GPi implants, voltage thresholds for dysarthria and from Starr PA. Placement of deep brain stimulators into the subthalamic facial contraction indicate the proximity to the corticobulbar nucleus or globus pallidus internus: technical approach. Stereotact tract, posterior and medial to the target. The activation of the Funct Neurosurg. 2002;79:118–145.) optic tract at low voltage may indicate that the lead is too deep, 736 SECTION 5 Functional Neurosurgery unless the lead has a “bulleted” tip (conducting surface includes the distal end of the lead). Corticobulbar tract activation at lower than 3 V or 2 mA may preclude therapeutic stimulation and dictates movement of the lead to a more anterior or lateral position. Diminution of contralateral arm rigidity is the most readily detectable acute effect of intraoperative test stimulation. Tremor and bradykinesia may improve acutely as well, but sometimes longer-term stimulation is required for full benefit to manifest. Motor symptom improvement without stimulation—the so-called microlesioning effect—may occur with the simple passing of the lead or microelectrode through the motor territory of the STN or GPi. Although such effect is variable, when present, it provides evidence of good electrode placement. Once macrostimulation is completed, the patient is put under anesthesia for the rest of the procedure. Closure and Pulse Generator Placement The lead is secured using a bur hole–mounted lead anchoring device (Stimloc, Medtronic, or Sure-Tek; Boston Scientific). We protect the end of the lead with a temporary cap and tunnel it sub- galealy to the area behind the pinna of the ear. The galea and skin are closed in the usual fashion. If available, intraoperative CT may be used at any point, and the images can be fused with preopera- tive targeting to assess the accuracy of lead placement. Lateral flu- oroscopy is useful to verify lack of lead movement as it is secured. After the removal of the frame and positioning of the patient similar to that for a ventriculoperitoneal shunt procedure, placement of either unilateral or bilateral lead extensions and internal pulse generators (IPGs) is performed under general anesthesia. A small incision is made over the protective cap of the lead. A 5-cm incision is made parallel and 2 cm inferior to the clavicle, and an IPG pocket is created over the pectoralis fascia. We tunnel the lead extender in a rostral-to-caudal direction to connect the lead to the IPG. The IPG is placed so as to avoid tension on the incision. The chest incision is closed in three layers. Placing the lead extender connector below the occipital ridge should be avoided because the mobility of the neck over time can cause fracture of the lead. Impedances are checked prior to closure of the incisions. ASLEEP INTERVENTIONAL MAGNETIC RESONANCE IMAGING SURGICAL TECHNIQUE Our technique uses a skull-mounted aiming device (SmartFrame; MRI Interventions) with dedicated targeting software (ClearPoint; MRI Interventions) in a 1.5-T diagnostic MRI scanner (Philips Achieva; Philips) with a 60-cm bore diameter. No preoperative stereotactic imaging is required. The patient is placed under general anesthesia, shaved, prepped, and injected with local anesthetic in a room adjacent to the scanner before being moved onto the MRI gurney. We rigidly fix the head in slight extension with a 4-pin carbon fiber head holder (Integra). Flexible surface receiving coils are placed in a sagittal fashion on both sides of the head, and the patient is moved through the MRI bore to allow surgical access at the back of the scanner. After a second prep, a custom drape is attached by elastic cords at both ends of the bore to maintain the sterile field during gantry move- ments. We use a skin-adhesive 7.5-cm × 9-cm MRI-visible mark- ing grid placed to span the coronal suture on each side during the first gadolinium-enhanced volumetric MRI scan for entry point selection. Once the images are imported into the targeting soft- ware, a selection of a preliminary target and trajectory is made as described earlier and correlated with the marking grid. After moving the head to the back of the bore, the intended entry is marked through the skin into the outer table of the skull with a sharp instrument.     CHAPTER 111 Deep Brain Stimulation for Parkinson Disease 737 For an intraoperative video of an iMRI case, please see point after DBS. Measures to reduce hardware erosion incidence Video 111.2. include using a countersinking bur hole technique, increasing 111 the IPG pocket size in the rostral-caudal direction to avoid having the IPG immediately under the incision, and performing RESULTS a multilayered closure to isolate the IPG from the incision. Motor outcomes of PD treatment are quantitatively measured Continued vigilance is important with respect to hardware using the UPDRS III. This is a clinician-scored motor evalua- erosion because an early diagnosis may allow a partial, rather than tion with high interrater reliability. Several level I studies have total, hardware explantation. shown that for patients with advanced PD, DBS at either the Lead or lead extension fracture occurrence in PD is reported STN or the GPi is more effective than best medical therapy in in approximately 2% of patients.76,77 Allowing for stress relief improving motor function, quality of life, and “on” time without loops and securing the lead/lead extension connection above the dyskinesias.39,69 occipital ridge with an extra fascial stitch may decrease the chance A 2005 study showed a 48% improvement in UPDRS III of fracture.73 Tunneling deeper in the subcutaneous tissue of the scores with bilateral STN stimulation compared with 39% neck may decrease the chances of “bowstringing” and scarring improvement with bilateral GPi stimulation.48 It also first of the extensions, which in turn may also decrease the fracture mentioned the pattern of cognitive and behavioral complications rate.78 The incidence of malpositioned electrodes requiring after STN implantation, not seen after pallidal surgery. A large surgical revision ranges from 0.8% to 1.7% per lead.74,76,77,79 trial with 159 patients completing 3-year follow-up confirmed the Repositioning as little as 2 mm toward the center of the motor near-equivalent improvement in motor scores with either STN territory of the target may significantly improve the efficacy of a (34% improvement) or GPi (30% improvement) stimulation.70 marginally placed lead. Again, some propensity for mildly increased depression and cognitive impairment outcomes with STN stimulation was seen. A randomized study of unilateral STN-DBS versus GPi-DBS FUTURE DIRECTIONS showed no difference in the motor score improvement, with mean Improvement in implantable devices is increasing our parameter percent improvement of 29.9% for the STN and 26.6% for the space in delivering stimulation. Constant voltage is being aug- GPi.47 Interesting to note, an analysis of experimental stimulation mented with constant current paradigms.80 Current steering, ventral to optimal settings in the STN and GPi showed that available with some of the newer lead models, will potentially stimulation in these locations caused adverse mood effects, an improve benefits and minimize side effects.81 Irregular stimula- outcome likely related to the electrical alteration of the parallel tion82 may also increase our ability to benefit patients by increas- nonmotor ventral limbic areas in both nuclei. A European study40 ing efficacy of treatment. Multisource asynchronous stimulation found no difference between the STN and GPi in cognitive and may be an efficient means to alleviate pathologic network syn- behavioral complications but did find significantly better results chronization, with clinical benefits that may outlast termination in medication reduction, off-medication motor scores, and mean of stimulation.83 change in Academic Medical Center Linear Disability Score with Neuromodulation implants with concurrent sensing and STN stimulation. stimulation ability (“bidirectional” interfaces) can aid our Follow-up of 8 to 10 years71 shows that, although DBS is understanding of pathophysiology and improve therapies for PD safe and efficacious over the long term, the nonmotor symptoms by incorporating feedback control into stimulation paradigms. progress, causing decrease in the function of the patient. “Adaptive” DBS, in which stimulation is modulated automatically according to changing brain needs, as detected by physiologic signatures of abnormal movement, may provide additional COMPLICATIONS benefits beyond standard continuous stimulation.84 Adaptive DBS The major complications of DBS for PD are hemorrhage, infec- paradigms for PD may now be implemented using permanently tion, and hardware failure. Complications can be separated implanted bidirectional interfaces.85 into early (up to 6 months) and late. A large review72 focusing on early (within 90 days) complications found that 132 of 1757 SUGGESTED READINGS patients (7.5%) had at least one such event. The most common Alexander GE, Crutcher MD, DeLong MR. Basal ganglia-thalamocorti- were wound infections (3.6%), pneumonia (2.3%), hemorrhage cal circuits: parallel substrates for motor, oculomotor, “prefrontal” and (1.4%), and pulmonary embolism (0.6%). Interesting to note, “limbic” functions. Prog Brain Res. 1990;85:119–146. patients older than 75 years did not have an increased rate of early Anderson VC, Burchiel KJ, Hogarth P, et al. Pallidal vs subthalamic postoperative complications. nucleus deep brain stimulation in Parkinson disease. Arch Neurol. The reported risk of symptomatic hemorrhage during MER- 2005;62(4):554–560. guided DBS in the literature ranges from 2.2% to 0.5%.43,44 Benabid AL, Pollak P, Louveau A, et al. Combined (thalamotomy and Hypertension and MER may increase the risk of cerebral stimulation) stereotactic surgery of the VIM thalamic nucleus for bilat- eral Parkinson disease. Appl Neurophysiol. 1987;50(1-6):344–346. hemorrhage in DBS.43,58 Possible ways of decreasing the risk Bergman H, Wichmann T, DeLong MR. Reversal of experimen- of hemorrhage include maintaining the mean arterial pressure tal parkinsonism by lesions of the subthalamic nucleus. Science. under 90 mm Hg, proper hydration to decrease the chance of 1990;249(4975):1436–1438. venous infarcts, and using trajectories that avoid blood vessels, Bronstein JM, Tagliati M, Alterman RL, et al. Deep brain stimulation for sulci, and the ventricle.73 Careful counseling of patients regarding Parkinson disease: an expert consensus and review of key issues. Arch this potential complication is imperative in the preoperative Neurol. 2011;68(2):165. discussion. Burchiel KJ, McCartney S, Lee A, et al. Accuracy of deep brain The incidence of hardware infection, defined as an infection stimulation electrode placement using intraoperative comput- requiring further surgery for partial or complete hardware ed tomography without microelectrode recording. J Neurosurg. 2013;119(2):301–306. removal, is reported to occur in up to 10% of cases.39,74,75 Charles D, Konrad PE, Neimat JS, et al. Subthalamic nucleus deep brain Hardware removal necessitates the abrupt cessation of stimulation, stimulation in early stage Parkinson’s disease. Parkinsonism Relat Disord. combined with a long course of antibiotics, and places emotional 2014;20(7):731–737. and economic strain on the patient and the treatment team. Cooper IS. Effect of anterior choroidal artery ligation on involun- Late complications include hardware erosion, lead fracture, tary movements and rigidity. Trans Am Neurol Assoc. 1953;3(78th and lead migration. Hardware erosion may present itself at any Meeting):6–7; discussion 8-9.